Subject: Re: [linux-audio-dev] Disk sampler
From: Benno Senoner (sbenno_AT_gardena.net)
Date: Wed Aug 30 2000 - 19:56:04 EEST
Hi,
No the project is not dead.
I haven't hacked much on the disksampler during summer, since I had to do paid
work.
Anyway me, David Olofson and Olli Niemitalo (one of the music-dsp guy),
had a long chat on IRC (no logs available :-( ) about how to optimize the event
system and the filter coefficient calculation in EVO.
BTW: I tried to load a 2GB in disksampler, 16 velocity zones and one sample per
key: works perfectly and "loads" (the first 128KB of each sample) in about
35seconds on my old IBM EIDE drive. Not that bad.
About the people that offered their help on the EVO project:
Engine:
Me: core engine
David Olofson: event system issues, engine optimizations)
Paul Kellet ( DSP stuff , interoperability with AKAI samplers, realworld issues)
Richard Smith and Mike Bailey: GUI , sample library and usability issues
Garth Brantley: GUI , importing of samples
Xavier Moulet: DSP stuff, GUI, instument creation tool/editor (synthetic
sounds rendered to a sample ?)
Olli Niemitalo said he will not be directly involved in the project, but will
give assistance when DSP or optimization related issues do arise.
If I missed something or someone or made incorrect statements please correct
me.
Or if others want to join us let us know. (please specify your area of
expertise)
<OFFTOPIC>
One reason because I had not much time for the audio stuff was because
I hacked on my Gnutella websearcher at
http://www.gnutella.it (in english, italian and german)
I wrote it because I was simply not satisfied with existing ones
(they almost all suck except gnute.com) and because it
was funny. I use MySQL to perform timestamped caching , plus perform
realtime searches in a separate window in order to give you full transparency
about the search results.
The engine is designed in a such way, that it avoids to stress the gnutellanet
too much by caching search keywords for a few mins. ( if the keyword "mpeg"
gets searched every second, a query for "mpeg" is issued only evey 5mins,
plus query results are routed to the right webclient using pattern matching
and lock-free FIFOs (who says that only audio apps need them ? :-) ) so that
you can send matching query hits to multiple clients without the need of
duplicate queries.
I make periodic statistics about the goodness of gnutella hosts (eg number of
query hits per hour, so that I can periodically discard the poorest ones.
PS: sorry for any selfpromoting/advertising, but I had to justify my absence
from the audio mailinglists :-)
</OFFTOPIC>
cheers,
Benno.
On Wed, 30 Aug 2000, you wrote:
> Hi,
>
> before I went on vacation this list was flooded with discussion about the
> disk-sampler. But when I came back all was silent, and I had received almost
> nothing on the subject during my vacation.
> Is there still depelopment going
> on? Whats the status? Or has a separate list been formed? I'm asking because
> I would like to participate in its development. Or, if it died, I will start
> working on my own design.
>
> Ruben
This archive was generated by hypermail 2b28 : Wed Aug 30 2000 - 21:19:52 EEST